We are unable to sponsor or take over sponsorship of an employment visa at this time
AI Developer / Agentic AI Engineer
Role Summary
We are building an agentic AI platform to transform commercial banking customer service. The AI developer will design, build, and operate LLM-powered agents that interpret inbound servicing requests (e.g., email / case intake), retrieve grounded knowledge, and execute approved workflows through secure tool/API integrations - with enterprise-grade controls, observability, and human-in-the-loop patterns.
This role sits within a cross-functional team with Product, Operations, Technology, and Risk partners and focuses on delivering production-ready agentic AI capabilities for regulated financial services
Responsibilities
Agentic AI Solution Development
- Build and enhance LLM/agent orchestration (Planner/supervisor patterns, tool-using agents, routing, guardrails).
- Implement intent classification information extraction validation and decision logic for servicing workflows
- Developed tool calling integrations to downstream systems (CRM, workflow engine, core banking services, case management)
- Implement human-in-the-loop workflows (review, approval, escalation, override) based on confidence/risk thresholds
Knowledge and grounding (RAG)
- Design and implement retrieval-augmented generation (RAG) for policy procedure grounding and resolution guidance
- Build knowledge ingestion pipelines with refresh/versioning
- Improve answer quality via chunking strategies, embeddings re ranking and context management
Quality, Safety and Evaluation
- Define and run evaluation frameworks: golden datasets, scenario tests, regression tests, and automated scoring.
- Reduce hallucinations and risk by implementing prompt policies, constraints, structured outputs, and verification steps.
- Partner with risk slash compliance to ensure traceability, audit logs, explain ability requirements are met.
Production Readiness and Operations
- Implement observability for agents (latency, cost, tool failures, drift, quality signals, escalation rates).
- Support CI/CD for agent prompts and configurations (versioning, approvals, rollback).
- Collaborate with platform and security teams on secrets management, access controls, PII protections, and safe deployments.
Required Qualifications
- 4+ years of software engineering experience or equivalent with strong CS fundamentals
- Hands-on experience building with LLMs and modern AI app stack (agents, RAG, tool/function calling).
- Strong proficiency in Python and building back-end services/APIs.
- Experience with at least one: LangChain / LangGraph, Llamalndex, Semantic Kernel or equivalent frameworks.
- Experience with vector databases and search (e.g., Pinecone, Weaviate, Milvus, OpenSearch/Elastic, pgvector)
- Experience deploying services in cloud environments (AWS/Azure/GCP) with basic DevOps practices
- Strong understanding of security and privacy principles (PII handling, least privilege, audit logging)
Preferred Qualifications
- Experience in financial services or other regulated domains (risk controls, compliance audit readiness)
- Experience integrating with enterprise workflows (e.g., ServiceNow, Custom workflow engines, BPM/RPA)
- Familiarity with model evaluation approaches (LLM-as-judge, rubric scoring, retrieval evals, offline/online testing)
- Experience with messaging/eventing (Kafka/SQS), email ingestion pipelines, and document processing
- Exposure to MRM concerns and governance (model cards, risk assessments, validation processes)
Kaleidoscope, an Infosys Company, is an equal opportunity employer, and all qualified applicants will receive consideration without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, protected veteran status, spouse of protected veteran, or disability.